More Bad News

Now that the second quarter is over, AMD announced that its revenues will be even lower than the lowest estimate they provided a couple weeks ago (our comments at that time can be seen here).

In all likelihood, this reflects those bargain basement prices you’ve been seeing on Pricewatch lately.

At the same time, Intel is pushing up its roadmap a little and we’ll see 2.8GHz PIVs announced late September, and perhaps we’ll officially see 3GHz by the end of the year.

Combine that with expected advances on the mobo front, and Intel looks well-positioned for the latter part of the year, especially for the Christmas season.

The problem with all this is not that AMD had a really bad quarter; it’s that the next few don’t look any better, at least not due to anything on AMD’s part. And now is not a good time to be doing disproportionately worse than your competition for any length of time; the market is in a hanging mood.

Hammer is a 2003 product, and that will take time to get going. If Barton were to show up by early October, that would at least be something for the Christmas season, but unless AMD has scrambled some more, Barton is supposed to be a UMC project, and UMC isn’t supposed to go into mass production until the beginning of 2003.

What About Cost Advantage?

Some might say AMD should be able to reduce costs with smaller .13 micron chips than Intel, but it does you no good to make a CPU $10 cheaper than your competitor if you get $50 less for it, and you can’t sell a ton of them for that low price.

Additionally, at least on the Athlon size, that tiny processor size is actually a bit of a mirage.

The Thoroughbreds and Clawhammers are about 50 sq. mm smaller than the (optically shrunk) PIV. However, most of that difference is due to Thoroughbred/Clawhammer having less cache than Northwood.

Cache takes up quite a bit of room on a CPU. Turn a Thoroughbred into a Barton, and most of the size differential between the two goes away. The same will eventually happen to Clawhammer, too.

It’s really a specious argument, though. So long as Intel can command 50-100% more money for a processor rated the same speed as an AMD processor, cost advantage is a competitive non-issue. It’s like Ford trying to beat Mercedes with price.

The Near-Term View

It will probably take a while for Palomino inventory to be cleared out, so that will be an anchor on the early part of next quarter’s revenues.

After that, a big question will be, “Can AMD charge more for Tbreds than the same speed Palominos?” If they can, that will help, but if they can’t, they’re locked into a loss-making price structure the rest of the year.

This question will largely turn on demand, and computer hardware makers are hoping against hope that the third quarter will turn things around. AMD was disproportionately affected by a slack second quarter. If demand does pick up third quarter, they could disproportionately recover (especially if Intel has shortages).

That’s essentially AMD’s single hope for 2002. They may get it, in which case life will be much less stressful in Sunnydale. The nightmare is if they don’t.

Are We Talking Survival Here?

The worse the financial news is in 2002, the more pressure there will be for Hammer to succeed, and succeed quickly in 2003. If it is, then these quarters won’t be too big a deal.

But if AMD runs three straight $150 million losses for the rest of the year, then Hammer stumbles and AMD keeps on bleeding at that pace well into 2003, AMD could well be in serious trouble a year or a bit more from now. A performance like that would chew up most of their current financial reserves, and in the current environment, it would be tough for them to get more.

In a more friendly financial environment, this situation wouldn’t be so bad, but investors aren’t exactly in a nurturing mood at the moment. They’re looking for blood. God help AMD if any accounting irregularities come to light in today’s environment.

If I had to bet on this, I’d bet they’d muddle through, likely with clipped wings, but there is cause for concern.

Ed

Be the first to comment

Leave a Reply